Real-time and historical bus data

• Using the public GTFS API provided by Auckland Transport, we obtain …
  • GPS positions (= position on a map), and
  • the current delay (= difference between scheduled and actual arrival time at last stop)
  … for all active buses in the network.

• Using data collected over the last 10 weeks, we can see if today is better or worse than usual

How many buses?

• Peak periods: A morning, B school, C evening
• The coloured areas represent the proportion
  of buses that are, from top to bottom,
  • late by more than 5 minutes
  • on time (within 5 minutes of schedule)
  • more than 5 minutes early

As of 8:00 pm today

• 152 of 207 buses were on time
• 27 were running more than 5 minutes late
• 28 were more than 5 minutes early

Today vs. last 10 Sundays

• number and percentage of buses early, ontime, or late
• comparing average (dashed) of past Thursdays with today

This week vs. last 4 weeks

• percentage of buses early, ontime, or late

So, just how late is late?

• Each dot represents one vehicle’s delay at 8:04 PM
• the bar contains 90%, 75%, and 50% of buses

Delays so far today

• rotating the bar from the previous graph, we see how the distribution of delays has changed today

Today vs. last 10 Sundays

• coloured lines represent the 5% and 95%, 12.5% and 87.5%, and 25% and 75% quantiles of bus delays

This week vs. last 4 weeks

• the regions contain 90%, 75%, and 50% of buses